Creating training images

[curt]

I am new to opencv and facial recognition. Is it reasonable to create training images with a webcam taking snapshots at regular intervals? Any suggestions such as face and camera angles, background, etc.

Thanks,
Curt

[berak]

@curt

Is it reasonable to create training images

with modern, dnn based classifiers, not at all.
just use the pretrained model, you won*t be able to improve it (triplet loss, anyone ?)

@curt

with a webcam taking snapshots at regular intervals?

for sure the worst idea i could imagine
have a [look here]( OpenCV: DNN-based face detection and recognition)